I, like many other people who enjoy fullrangers, know that a flat response is not critical for a good sounding speaker, but I have noticed that even small peaks in certain places can sound fairly offensive while large peaks in other places are barely noticeable.
Many drivers people find shouty for instance have peaks around 7-8kHz, and even a few dB peak here can sound very tiring. I've noticed someone mention that fr125s midrange can sound recessed due to a small peak around 2kHz.
Peaky bass on the other hand is usually not very noticeable. Room modes will mess up almost any bass response and yet it doesn't bother most people. I suppose, however that bass doesn't fall within what is referred to as the "critical range" of our hearing.
What this critical range is is also up for discussion but that has been discussed many times and is not really what I'm getting at.
I was mostly wondering if someone with a lot of experience with this could provide a list of the frequencies at which peaks are most offensive generally (nulls seem less noticeable than peaks). I realise this will be subjective, but hopefully it might at least serve as a guide.
Many drivers people find shouty for instance have peaks around 7-8kHz, and even a few dB peak here can sound very tiring. I've noticed someone mention that fr125s midrange can sound recessed due to a small peak around 2kHz.
Peaky bass on the other hand is usually not very noticeable. Room modes will mess up almost any bass response and yet it doesn't bother most people. I suppose, however that bass doesn't fall within what is referred to as the "critical range" of our hearing.
What this critical range is is also up for discussion but that has been discussed many times and is not really what I'm getting at.
I was mostly wondering if someone with a lot of experience with this could provide a list of the frequencies at which peaks are most offensive generally (nulls seem less noticeable than peaks). I realise this will be subjective, but hopefully it might at least serve as a guide.
I like this question. It brings to light the differeneces in ragged responses and good sounding speakers. I have found that no one frequency or set of frequencies is necessarily the problem as much as the relationship to other "critical" frequencies.
If you have a speaker that has a reference level at say 300hz, then a small peak at 7k, it will typically be referred to as shouty. If the same speaker has a null at 3000hz it will seem recessed. If the same speaker has a null at 300hz it may seem excessively bright. There isn't a level that this is easy to arbitrate. The most noticeable nulls to me have always been below 1000hz and the most noticeable peaks above 3500hz and below 9000hz. That's a wide range, but it has just been my experience in adjusting 100+ band equalization in very nearfield environments. (cars)
In a room with nodes, I would have to think those numbers change very much. I am guessing that second and third order harmonics from those nodes would play a big part in what is noticeable. In anechoic conditions, the result may be very different.
I am interested to see some of our very respected professionals (of which I am very far from being) keying in on what is more prevelant in real world conditions. I guess this means my two cents worth would be (in room) best to avoid peaks between 3500-9000 hz. That's way too broad for an answer though.
Hoping for more input from several specific people
Robert
If you have a speaker that has a reference level at say 300hz, then a small peak at 7k, it will typically be referred to as shouty. If the same speaker has a null at 3000hz it will seem recessed. If the same speaker has a null at 300hz it may seem excessively bright. There isn't a level that this is easy to arbitrate. The most noticeable nulls to me have always been below 1000hz and the most noticeable peaks above 3500hz and below 9000hz. That's a wide range, but it has just been my experience in adjusting 100+ band equalization in very nearfield environments. (cars)
In a room with nodes, I would have to think those numbers change very much. I am guessing that second and third order harmonics from those nodes would play a big part in what is noticeable. In anechoic conditions, the result may be very different.
I am interested to see some of our very respected professionals (of which I am very far from being) keying in on what is more prevelant in real world conditions. I guess this means my two cents worth would be (in room) best to avoid peaks between 3500-9000 hz. That's way too broad for an answer though.
Hoping for more input from several specific people
Robert
I think you're probably right about the relationships of frequencies being important, but I also think that for the most part drivers share a great deal in common which allows for fairly good generalisations.
The 7-8khz peak for instance will make most drivers sound shouty and while it might not do so for all drivers, its a fairly good generalisation to say that it does.
Another one i've found is that even small 2-3dB peaks around 500Hz make speakers sound very bright. These frequencies are more affected by ripple and room interaction, however, making it more difficult to generalise about the sound of a driver with a peak here as it could depend as much on implementation and placing as the driver itself
Small peaks around 5-6kHz seem less bothersome than 2-4 or 7-8, but I'm not too sure about this and would need input from those more experienced than I before committing to it.
It might be nice if after some discussion I was able to draw up a little colour continuum type chart designating in very general terms how noticeable response anomalies are across the frequency range.
The 7-8khz peak for instance will make most drivers sound shouty and while it might not do so for all drivers, its a fairly good generalisation to say that it does.
Another one i've found is that even small 2-3dB peaks around 500Hz make speakers sound very bright. These frequencies are more affected by ripple and room interaction, however, making it more difficult to generalise about the sound of a driver with a peak here as it could depend as much on implementation and placing as the driver itself
Small peaks around 5-6kHz seem less bothersome than 2-4 or 7-8, but I'm not too sure about this and would need input from those more experienced than I before committing to it.
It might be nice if after some discussion I was able to draw up a little colour continuum type chart designating in very general terms how noticeable response anomalies are across the frequency range.
Actually, now you mention it, I would say that the most easily audible peaks I have experienced have been around 3K. I always thought it was due to cone size and efficiency of amplification in that range. Almost every adjustment I have ever done has included lowering the 3k-3500 region. Of course, this is a very poplular crossover point too. I always wondered if the components involved with that had something to do with the peaks?
I always thought of the 500hz region peaks as causing a harsh sound. Usually a peak there seemed to produce some scratch in the voices etc. I could be wrong about that one though.
I always thought of the 500hz region peaks as causing a harsh sound. Usually a peak there seemed to produce some scratch in the voices etc. I could be wrong about that one though.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.